Retail Media Glossary

Showing two versions of an ad, listing or bid setting to comparable shoppers and keeping the one that performs measurably better.

Ad spend divided by the sales those ads generated, as a percentage. A 25% ACOS equals a 4.0x ROAS; it is Amazon's standard efficiency metric.

The rules that decide which ad gets credit for a sale, such as last click or a view within 14 days, and so shape reported ROAS.

A set of ads that share a goal, budget, schedule and targeting, such as a sponsored products campaign for one product line on Walmart.

Too many ads competing on one page or search result, which lowers the attention and click-through each of them gets.

A marketplace where publishers' ad impressions are auctioned to advertisers' buying platforms in real time.

A subset of a campaign that bundles products, keywords or audiences with shared bids, used to control spend at a finer level.

One display of an ad to a shopper; the unit behind CPM pricing and the starting point for viewability and reach.

The ad placements a retailer or publisher has available to sell, such as search slots, product pages and banners.

A company that aggregates inventory from many publishers and sells it to advertisers as one package.

The day-to-day work of setting up, monitoring and fixing campaigns; much of it is what automation takes off a lean team's plate.

How well ads meet their goal, judged on metrics such as ROAS, ACOS, conversion rate and incremental sales.

Technology that decides which ad to show in a placement, delivers it and records the impressions and clicks.

A snippet of code on a page that calls an ad server to load an ad and report what happened.

The software used to buy, sell, deliver and measure digital ads, from DSPs and SSPs to ad servers and analytics.

Independent checks that ads were seen by real people, in brand-safe places and in the intended market.

The connections that let software such as RMIQ read reports and change bids on each retail media network automatically.

Pricing where advertisers bid for each placement, so the auction sets what each click or impression costs.

Buying ads by who the shopper is, using retailer purchase data, as opposed to buying a page or keyword.

Showing ads to shopper segments defined by behaviour, purchase history or demographics, such as lapsed buyers of a category.

Letting an algorithm set bids toward a goal, or setting every bid by hand: the core trade-off between control and scale.

Revenue per order: total sales divided by the number of orders. Bundles and upsell ads are common ways to raise it.

Studying which products shoppers buy together, to plan cross-sell ads and measure halo sales.

Groups of shoppers defined by what they do, such as browsing a category or buying on a schedule, used for targeting.

Patterns in how advertisers bid over hours, days and competitive moments, which shape auction prices and win rates.

The message an ad slot sends to buyers, describing the placement and the shopper context and inviting them to bid.

A buyer's answer to a bid request: the bid price and the ad to show if it wins.

The rules for how much to bid and when, for example target ROAS, maximum clicks or dayparting.

A list of sites, apps, keywords or placements where ads must never appear, used for brand safety and to cut wasted spend.

Methods that identify non-human traffic so fake clicks and impressions are filtered out before they are billed.

A set of hijacked devices controlled together to generate fake ad clicks and impressions at scale.

Visits, impressions or clicks generated by software; left unfiltered, they inflate costs and distort results.

Building long-term awareness and preference for a brand. On retail media it usually runs alongside sales-focused campaigns.

Keeping ads away from content or contexts that could damage the brand's reputation.

Deciding how spend is split across retailers, campaigns and products, ideally moving money to where it earns the best return.

Spreading a budget across the flight so it neither runs out early nor goes unspent, while still winning the best auctions.

Continuously adjusting bids, keywords, budgets and creative to improve a campaign against its goal.

How a brand's products and ads perform within a retail category, including share, conversion and growth against competitors.

The plan for which retailers and ad types to use and in what mix, so each channel plays a clear role.

Linking ad exposure to actual purchases through the retailer's own sales data, which is what makes retail media so measurable.

Placing ads by the page, search term or category being viewed, with no need to know who the shopper is.

Shoppers deliberately left unexposed to an ad, whose behaviour shows what would have happened anyway and so isolates its incremental impact.

The share of clicks or visits that end in a purchase. On retail media, price, reviews and the product page shape it as much as the ad.

Targeting and measuring ads without third-party cookies, using first-party and contextual signals, where retail media is strongest.

Matching one platform's cookie IDs to another's so ad systems can recognise the same user across them.

Ad spend divided by the purchases or other conversions it produced.

What an advertiser pays each time a shopper clicks an ad; the main pricing model for sponsored search.

The price paid each time a video ad is watched to the end.

The price of 1,000 ad impressions: the standard pricing unit for display and video.

The price paid each time a video ad is viewed, by the platform's own definition of a view.

Promoting related or premium products to shoppers who are already buying: a common retail media tactic to grow basket size.

Clicks divided by impressions: a measure of how relevant and compelling an ad is to the shoppers who see it.

Total marketing spend divided by the number of new customers won: a benchmark for what a new shopper is worth paying for.

Software that unifies a brand's customer data from many sources into single profiles for targeting and measurement.

The total revenue or profit a customer brings over the whole relationship, used to justify acquisition spend above the first order.

Uploading a brand's own customer list so a platform can target those shoppers or exclude them.

A company that collects data from many sources and packages it for advertisers and analysts.

Software that collects audience segments, mostly cookie-based, for ad targeting.

A supplier of audience or market data used to target and measure advertising.

The buy side of programmatic advertising, where advertisers and agencies source inventory through DSPs.

Software advertisers use to buy impressions across many publishers in real time, including retailers' offsite inventory.

Banner and visual ads on websites and apps; in retail media, onsite banners plus offsite ads using the retailer's data.

Fraud that disguises low-quality inventory as a premium site or app to charge higher prices.

Technology that assembles ad creative on the fly, swapping product, price or message for each shopper.

Changing product prices with demand, competition or stock; it moves conversion rates and so what an ad click is worth.

Ads generated automatically from a product catalogue, showing each shopper items they viewed or are likely to want.

Revenue or cost per 1,000 impressions whatever the pricing model, used to compare CPC and CPM buys like for like.

The share of ad requests actually filled with an ad; low fill means unsold inventory for the publisher.

A deal that lets a preferred buyer see and bid on inventory before it goes to the open auction.

Data a company collects directly from its own customers. A retailer's purchase data is what makes retail media targeting precise.

An auction where the winner pays exactly what it bid, which makes bid strategy and bid shading matter more.

Buying placements at an agreed rate outside the auction, common for homepage takeovers and guaranteed deals.

The minimum bid a seller will accept for an impression or click; bids below it lose automatically.

Limiting how many times one shopper sees the same ad, to avoid wasted impressions and annoyance.

Easily identified non-human traffic, such as known crawlers and data-centre bots, removed by standard filters.

Showing ads only to shoppers in chosen locations, such as the regions where a product is stocked.

Letting several exchanges bid on an impression at once before the ad server decides, to raise the publisher's yield.

Ads shown inside mobile apps, including retailers' own shopping apps, where much retail media inventory sits.

The sales an ad truly caused, beyond what would have happened without it: the most honest measure of retail media value.

Experiments with test and control groups that measure how many sales an ad actually added.

Video ads that play before, during or after other video content, such as pre-roll on a retailer's streaming partner.

A full-screen ad shown between screens of an app or website, such as between steps of a shopping flow.

Any impressions or clicks that do not come from genuine shoppers, from bots to accidental clicks, and should not be paid for.

Showing sponsored ads when shoppers search specific terms on a retailer's site: the backbone of sponsored search.

Measurements comparing exposed and unexposed shoppers to show how much an ad raised sales or awareness.

New shoppers who resemble a brand's existing buyers, used to reach likely customers at scale.

The modelling behind lookalike audiences: finding shoppers whose behaviour resembles a brand's best customers.

Software that clicks ads to drain competitors' budgets or earn fraudulent payouts.

Ads sold by online marketplaces such as Amazon or Walmart Marketplace to the sellers who list products there.

A brand's share of total category sales. Retail media is often how brands defend or grow it on each retailer.

Statistical analysis of how much each channel, retail media included, contributes to total sales.

Sharing credit for a sale across every ad a shopper encountered along the way, from the first view to the last click.

Coordinating ads across online, offsite and in-store touchpoints so a shopper meets one consistent message.

Onsite ads run on the retailer's own site and app; offsite ads use the retailer's data to reach its shoppers elsewhere on the web.

TV content streamed over the internet. Retailers increasingly sell OTT ads targeted with their shopper data.

Video ads that play in a placement of their own within articles or feeds, outside other video content.

How long it takes for the profit from newly acquired customers to repay what it cost to acquire them.

Marketing bought and judged on measurable outcomes such as sales or ROAS; sponsored search on retail media is a prime example.

Crediting sales to ads after the purchase has happened, usually by matching orders to ad exposure.

A programmatic deal giving one buyer a fixed price on inventory before it reaches the open auction.

Tracking competitors' prices for the same or similar products, since a price gap can undo even the best ad.

Advertising built on consented first-party data and aggregated reporting, with no tracking of individuals across sites.

An invitation-only auction where a publisher offers selected inventory to chosen buyers.

The retailer's page for a single product, where many sponsored ads appear and where the sale is won or lost.

The profit left on each sale after costs; it sets how much ACOS or CPA a product can afford.

Where a product appears in a retailer's search results. Sponsored ads buy visibility, and the sales they drive lift organic rank.

Buying and selling ads automatically through software and real-time auctions, with no manual insertion orders.

Any owner of media that sells ad space. In retail media, the retailer itself becomes the publisher.

The number of unique shoppers who saw an ad at least once during a campaign.

Auctioning each ad impression in the milliseconds while a page loads.

How retailers turn their traffic and shopper data into advertising revenue, from sponsored listings to offsite and in-store ads.

An advertising business run by a retailer, selling ads on its site, app, stores and beyond, targeted and measured with its own shopper data.

Showing ads to shoppers who viewed or added a product to the cart and left without buying.

Interactive ads with video, animation or shoppable elements that invite engagement beyond a click.

Ad-attributed sales divided by ad spend. $5 of sales on $1 of spend is a 5.0x ROAS.

Ads shown in search results when shoppers enter a query; on retailers' sites these are sponsored products and sponsored brands.

Campaigns planned around peak shopping moments such as Prime Day, back to school or the holidays, when auctions are most competitive.

Another company's first-party data shared directly under agreement, such as a retailer sharing audiences with a brand partner.

Recording conversions from a company's own server; more reliable than browser tracking as cookies and trackers are blocked.

A brand's share of all searches in its category: a leading indicator of future market share.

A brand's share of all ad impressions or sponsored placements in its category.

Campaigns that show product listings with image and price, such as Google Shopping or sponsored products on a retailer.

Audience segments built automatically by a platform's algorithms to find shoppers likely to convert.

Fraud that imitates real shoppers closely and takes advanced analysis to detect.

Search ads that feature a brand's logo, headline and several products, usually at the top of retail search results.

Confidence that a test result reflects a real difference and not chance, which a test should reach before anyone acts on it.

The sell side of programmatic advertising, where publishers and retailers offer inventory through SSPs and exchanges.

Software publishers use to sell their ad inventory to many buyers at once.

Groups of people or devices paid to visit sites and click ads in bulk, generating traffic with no genuine interest.

Choosing who sees an ad, by keyword, product, category, audience or location.

Data bought from companies with no direct relationship with the shopper; less reliable and fading as privacy rules tighten.

Small pieces of code that record page views, clicks and conversions for measurement and attribution.

An agency or in-house team that runs programmatic buying across DSPs on behalf of advertisers.

Winning new customers or app users, and the spend and tactics dedicated to it.

Fraud in which bots fake the browser or device details they report so they pass as real users.

Reviews, photos and videos made by shoppers, which strongly influence whether an ad click turns into a sale.

Whether an ad had a real chance to be seen. The standard is half its pixels on screen for one second, two for video.

A closed platform, such as Amazon, that keeps its data and ad tools in-house and limits outside measurement.

The share of the auctions an advertiser enters that it wins; too low means lost visibility, too high can mean overbidding.

A seller's efforts to earn the most from each impression through pricing, floors and demand sources.